[Python] Cookbook

• Import file in a different/common-used folder

import sys
sys.path.insert(1, '/path/to/application/app/folder')
import file

importing-files-from-different-folder

• Run system commands without using os.system()

import subprocesssubprocess.call(["ls", "-lt"])

This command calls the shell script and returns the system output and return code attribute (e.g., 0 for no error, 1 for error). This will wait for the completed process.

• Use makedirs instead of mkdir

os.makedirs(path) works like os.mkdir, except that any intermediate path segment (not just the rightmost) will be created if it does not exist. 

However, of the target directory already exists, raise an OSError if exist_ok is False. Otherwise no exception is raised.

So it is better to check if the directory exists first:
if not os.path.exists(dir):
    os.makedirs(dir)

• Open a file without closing it

Opening a file using with is as simple as: with open(filename) as file:

with open("test.txt") as file: # Use file to refer to the file object

  file.write("hello world!")  # Opens file in write mode

Notice, that we didn't have to write "file.close()". That will automatically be called.

• Delete a file

os.remove(/path/to/file)

• Numpy calculation handling exceptions

# out: the output is an array with the same size as area_left_in_pixel.
# where: broadcast at locations where the condition is true, elsewhere the out array will retain its original value.

# single formula: percent = area/max_area

percent = np.divide(area, max_area, out=np.zeros_like(area), where=max_area!=0)

• Find the indices of the largest value

Indexes of the maximal elements of a N-dimensional array:

ind = np.unravel_index(np.argmax(a, axis=None), a.shape)

• Merge multiple dataframes

from functools import reduce

coef_full = reduce(lambda left,right: pd.merge(left,right,how='left'), [llp_df, slp_df, vslp_df])

• Multiply a constant/column to the numeric values of an existing dataframe

The first one is preferred because it keeps the original dataframe.
df[df.select_dtypes(include=['number']).columns] *= constant

df.select_dtypes(exclude=['object', 'datetime']) * 3